Lucas Sloan
Lucas Sloan
tf.keras allows you to define models with multiple inputs and outputs, but in order to use the standard keras training loops, you [need to provide](https://www.tensorflow.org/guide/keras/functional#models_with_multiple_inputs_and_outputs) a tf.Dataset which returns either...
The scale is being calculated [here](https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/unet.py#L348) as `1 / math.sqrt(math.sqrt(ch))`. The comment says it was adapted from the attention implementation [here](https://github.com/hojonathanho/diffusion/blob/master/diffusion_tf/models/unet.py#L74), where the scale is `int(C) ** (-0.5)`, which is...
I based this on the in progress sampling in https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/train_dreambooth_lora.py due to the suggestion on https://github.com/huggingface/diffusers/pull/2030 that that was a good example to follow. Unfortunately, this code doesn't work at...
Runs the image through the model twice, and does nms on the combined detections. See #5730
I saw 2 instances of this error, there might be more. Basically, the Closure type system has much broader base types than the TypeScript type system. So in Closure, the...
A goog.provide file: ```javascript goog.provide('goog.example'); /** @constructor */ goog.example = function() {}; ``` Produces this .d.ts: ```typescript declare namespace ಠ_ಠ.clutz.goog { class example extends example_Instance { } class example_Instance {...